Just as the modern electronic age has changed the way we study chemistry, it has also changed the way we perform statistical analysis. In the old “wet chemistry” days, we worked for hours to produce a few measurements. Today we use electronic instruments that spew out data. Some of the old tried-and-true statistical methods are still useful, however, especially when combined with modern rapid analysis and good graphics. Newer methods are useful and necessary as well, but problems can arise with interpretation of results. Do you know what the first principal component and a latent variable are? Do you know what assumptions you are making about your data? Are you concerned about multicollinearity or heteroscedasticity? Many problems can be avoided with careful planning. Two of the biggest problems we have in research and quality assurance are 1) sampling and 2) lack of a clear objective. I recommend you think through your experiment or trial before you start. (Do not be afraid to run your trial in stages.) Where, when, and how will you gather your samples? What will you do with them to prepare for analysis? What, exactly, do you plan to measure, and how would you like to present your results?